How can I Identify the candidates of directly regulated genes by a transcription factor (TF) using DEGs list in KO mutant of that TF without ChIP-seq data ?

Question

[threepointking824]

Dear specialists of bioinformatics,

I would like to identify candidate genes directly regulated by a transcription factor (TF) from the list of differentially expressed genes (DEGs) of a TF-Knockout (KO) mutant without using ChIP-seq data.

What I did; To find downstream targets of TF-A (a gene of interest) of tomato, I made a knockout line of TF-A gene and conducted RNA-seq using total RNA sample from fruit.

Then, I obtained several hundreds of DEGs (Up and Down) by comparing a RNA-seq data of wild type and the TF-A-KO line.

What I can do now; I know the canonical binding motifs of TF-A. To identify the direct targets* of TF-A from the DEGs, of course, I can check the presence of TF-A binding motifs in the promoter region of all DEGs one by one. But, it's very time-consuming.

*direct target; here, it means the gene that TF-A binds to the promoter region and modulate its mRNA expression.

What I want to know; Are there any web tools or programs to identify the DEGs which have the specific TF-binding site in their upstream region (like 2kb from TSS) from the DEGs list ? What kind of approach can I take except the one-by-one method mentioned above ?

I'm using a mac, I have few experience of programming. I just started studying the basic of R.

Thank you for sharing your knowledge

Best,

Answer 1

[Devon Ryan]

Since you have the binding motif, just identify all instances of it in the genome, make a BED file of that, and intersect it with a BED file of your DE genes (change the coordinates to include the promoter region and maybe the first couple hundred bases of the gene body). Then you have what you want.

There are many posts on biostars about finding all instances of a motif (the first one I came across is find motif locations in the genome ).

Answer 2

[threepointking824]

Dear Dr. Devon Ryan,

Thank you for your helpful comment. I appreciate it. Although I have never made BED file by myself, I could understood your idea. I'll learn how to make BED files and how to handle it with my computer (bedtool?).

I found a web tool called RSAT (http://rsat.sb-roscoff.fr/). Probably, I can use this tool as well once I could get promoter seq of all candidates.

Best,

Answer 3

[jrt]

Agree with the binding motif analysis but keep in mind that approach is subject to high false positives arising from short degenerate motifs.

Do you have any known agonists of said transcription factor? Looking at +/- transcription activator treatment in the wt and KO, you could ask a more complex and specific question: what genes are both responsive to specific activating treatment in wt and the response ablated in the KO?

[I just started studying the basic of R.]

That's the answer to the one by one tedium problem. R is a good starting point. I haven't done much TF analysis since I've been working in R, but I do know the MSigDB pathway package has a collection (C3:TFT; http://software.broadinstitute.org/gsea/msigdb/collections.jsp) of genelists containing TF motifs in promoter regions.